| *** yamamoto has joined #openstack-dragonflow | 01:16 | |
| openstackgerrit | liuhaixia proposed openstack/dragonflow: Add Vlan Networking https://review.openstack.org/288270 | 01:26 |
|---|---|---|
| openstackgerrit | liuhaixia proposed openstack/dragonflow: Add Vlan Networking https://review.openstack.org/288270 | 01:26 |
| openstackgerrit | liuhaixia proposed openstack/dragonflow: Add Vlan Networking https://review.openstack.org/288270 | 01:27 |
| *** yamamoto has quit IRC | 01:28 | |
| openstackgerrit | hujie proposed openstack/dragonflow: add ovsdb monitor fullstack test case https://review.openstack.org/296905 | 01:53 |
| openstackgerrit | Li Ma proposed openstack/dragonflow: Delete stale df db objects when neutron db transaction failed https://review.openstack.org/297519 | 02:28 |
| *** yamamoto has joined #openstack-dragonflow | 02:29 | |
| openstackgerrit | hujie proposed openstack/dragonflow: implement test share for dragonflow https://review.openstack.org/298079 | 02:32 |
| openstackgerrit | hujie proposed openstack/dragonflow: add a new test log for dragonflow https://review.openstack.org/298079 | 02:46 |
| openstackgerrit | hujie proposed openstack/dragonflow: add a new test log for dragonflow https://review.openstack.org/298079 | 02:54 |
| openstackgerrit | feipeng proposed openstack/dragonflow: Add Redis Devstack https://review.openstack.org/297916 | 03:08 |
| openstackgerrit | feipeng proposed openstack/dragonflow: Insert the description of the change. https://review.openstack.org/298088 | 03:18 |
| openstackgerrit | feipeng proposed openstack/dragonflow: Add Redis DB Driver and Redis PubSub Driver https://review.openstack.org/286028 | 03:43 |
| gsagie | nick-ma: are design summit schedule is published anywhere? | 05:58 |
| gsagie | nick-ma: havent seen or got any email | 05:58 |
| nick-ma | gsagie: ok, I will visit some partners from intel and rack space during design summit. So, I need to schedule in advance to prevent from conflict. | 06:00 |
| gsagie | nick-ma: np, will update you as soon as i see something, i am checking now | 06:00 |
| nick-ma | gsagie: thanks. | 06:02 |
| *** gampel1 has joined #openstack-dragonflow | 06:14 | |
| openstackgerrit | Eran Gampel proposed openstack/dragonflow: DHCP DOS prevention https://review.openstack.org/296517 | 06:26 |
| *** oanson has joined #openstack-dragonflow | 06:47 | |
| *** gongysh has joined #openstack-dragonflow | 06:53 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Lookup router_port's network_id with lswitch's topic https://review.openstack.org/298121 | 06:56 |
| openstackgerrit | Li Ma proposed openstack/dragonflow: Delete stale df db objects when neutron db transaction failed https://review.openstack.org/297519 | 07:06 |
| openstackgerrit | Yuli proposed openstack/dragonflow: Fix DHCP fullstack tests https://review.openstack.org/298131 | 07:18 |
| *** yuli_s has joined #openstack-dragonflow | 07:42 | |
| yuli_s | Hello all | 07:43 |
| *** Shlomo_N has joined #openstack-dragonflow | 07:52 | |
| gsagie | Meeting in #openstack-meeting-4 in 2 minutes | 07:58 |
| *** Fei has joined #openstack-dragonflow | 07:59 | |
| *** Fei has quit IRC | 07:59 | |
| gsagie | meeting start now :) | 08:00 |
| gsagie | dingboopt | 08:01 |
| gsagie | DuanKebo ? | 08:01 |
| openstackgerrit | Yuli proposed openstack/dragonflow: Fix DHCP fullstack tests https://review.openstack.org/298131 | 08:06 |
| *** Fei has joined #openstack-dragonflow | 08:49 | |
| gampel1 | guys if you like to do it earlier jsut let us know | 08:57 |
| openstackgerrit | hujie proposed openstack/dragonflow: fix the ovsdb monitor connect to ovsdb fail bug https://review.openstack.org/298167 | 09:26 |
| gsagie | yes, earlier is better :) | 09:35 |
| *** feipeng has joined #openstack-dragonflow | 09:42 | |
| gsagie | feipeng: here? | 09:53 |
| gsagie | maybe ask DuanKebo if we can do the meeting earlier :) | 09:53 |
| *** feipeng has quit IRC | 10:24 | |
| openstackgerrit | Fei Rao proposed openstack/dragonflow: The implementation of distributed-DNAT https://review.openstack.org/289148 | 10:31 |
| *** yamamoto has quit IRC | 10:59 | |
| *** DuanKebo has joined #openstack-dragonflow | 11:22 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Install OVS as an OS service https://review.openstack.org/296897 | 11:24 |
| Fei | @gsgie, when will we start the meeting? | 11:32 |
| DuanKebo | Hi raofei | 11:33 |
| oanson | Hi | 11:33 |
| Fei | Hi everyone | 11:33 |
| nick-ma | Hi | 11:34 |
| *** hshan has joined #openstack-dragonflow | 11:37 | |
| gsagie | we need gampel here | 11:38 |
| gsagie | gampel? | 11:39 |
| gampel1 | hi | 11:41 |
| gampel1 | i am here | 11:41 |
| gampel1 | gal ? | 11:42 |
| gampel1 | lets start | 11:42 |
| gampel1 | hi everyone | 11:43 |
| oanson | Hi | 11:44 |
| nick-ma | hi | 11:44 |
| gampel1 | lets start sorry for the delay | 11:44 |
| gsagie | start | 11:44 |
| gsagie | i am here | 11:44 |
| gampel1 | we talked today about feature freeze and code freeze | 11:44 |
| gampel1 | Gal suggested end of next week for the feature freeze | 11:45 |
| gampel1 | and we will try to push SG, DNAT, redis, Controller reliability | 11:45 |
| gampel1 | and code freeze on the 18th right ? | 11:46 |
| Fei | OK. it will not impact the bug fix for merged features, right? | 11:46 |
| oanson | Bug fixes are between feature freeze and code freeze. | 11:47 |
| nick-ma | got it. the proposed date seems reasonable. | 11:47 |
| gampel1 | We want ti cat the stable branch on the 18th | 11:47 |
| Fei | ok | 11:47 |
| gampel1 | Fei: can you update about the DNAT status | 11:47 |
| Fei | yes, I think DNAT is almost completed. and @mic-ma raise some suggestion. | 11:48 |
| DuanKebo | When will it reopen to new features again? | 11:48 |
| gampel1 | after we have the stable branch | 11:49 |
| gsagie | DuanKebo: we just create stable branch/tag and then continue | 11:49 |
| gsagie | then we need to backport bug fixes to it | 11:49 |
| Fei | I will raise a new patch to implement his suggestion. @nick-ma, do you agree with me:) | 11:49 |
| DuanKebo | OK, can we still summit some test code? | 11:50 |
| nick-ma | yes. | 11:50 |
| gampel1 | whats the status of SG and Redis | 11:50 |
| gsagie | DuanKebo: test code is fine | 11:50 |
| DuanKebo | OK | 11:50 |
| gampel1 | gal tested SG yesterday and got allot of exceptions | 11:50 |
| gsagie | it is just so we can have a "semi stable" version that we can start running bigger tests instead of always using master | 11:50 |
| gsagie | which is unstable | 11:50 |
| gsagie | yes dingboopt, i tried simple case of SG and got some exception, the reproduce steps in the review | 11:51 |
| gsagie | also the application is not enabled by default so the rules are not generated unless i added it to the conf | 11:51 |
| gsagie | but this is easy to fix :) | 11:51 |
| gampel1 | dingboopt: are you here ? | 11:52 |
| gsagie | DuanKebo: can someone from your team also test the security group patch | 11:53 |
| gsagie | and fix what is needed | 11:53 |
| Fei | It seems Yuanwei also join the SG develop | 11:53 |
| DuanKebo | Yuanwei is writing fustack/unit test code for sg. If necessary, he can work together with dingboopt | 11:53 |
| DuanKebo | to fix bugs of sg | 11:53 |
| gampel1 | Ok | 11:54 |
| Fei | for DNAT, the other update is DNAT test code have been completed. But since current DNAT path is too big, So I'd like to raise them on a new patch. | 11:54 |
| gampel1 | ok i have no problem with that @gal @nick-ma what do you think | 11:55 |
| Fei | Do you think it's ok or not? | 11:55 |
| gampel1 | feipeng:redis devstack -- why do we install it form source and not using an install package ? | 11:56 |
| DuanKebo | Feipeng is not here | 11:57 |
| oanson | I spoke with him earlier - apparently the redis provided by Ubuntu is too old. We need at least version 3. | 11:57 |
| gsagie | Fei: as long as we are ready to merge by next week | 11:57 |
| gampel1 | Ok i see omer did you try it | 11:57 |
| DuanKebo | Yes, only redis 3 support cluster | 11:57 |
| gampel1 | i see | 11:57 |
| Fei | ok, @Gsagie | 11:58 |
| gsagie | we need to be more quick on addressing review comments and we need to be quicker on the review | 11:58 |
| oanson | No, I haven't got to it yet. | 11:58 |
| oanson | Will be done today/tomorrow. | 11:58 |
| gampel1 | Ok so I think that feipeng address the comment in the review and we could merge | 11:58 |
| nick-ma | i'm ok. i haven't run redis, dnat and sg successfully and also gave some advices during review. i'll re-evaluate them tomorrow. also i'm struggling with redis3 cluster setup. most of the problems are related with the distribution and packages. i do need a working devstack script for that. | 11:58 |
| gampel1 | Do you want to discuss the controller reliability now or in another meeting ? | 11:59 |
| hshan | I think we should talk it now | 12:00 |
| gampel1 | ok | 12:00 |
| DuanKebo | code for redis is still not merged. | 12:00 |
| gampel1 | it failed the gate test | 12:00 |
| gampel1 | http://logs.openstack.org/28/286028/38/check/gate-dragonflow-python27/1c8b545/testr_results.html.gz | 12:01 |
| gampel1 | hshan: can you please update us on the open issues with controller reliability | 12:02 |
| hshan | yes | 12:02 |
| hshan | I've fixed all the issues | 12:02 |
| hshan | i'm writing fullstack test for reliability | 12:03 |
| gampel1 | I understand that we consider only using update from OVSDB and not from OpenFlow right ? | 12:03 |
| hshan | yes | 12:03 |
| hshan | I think we should only use update from OVSDB too | 12:04 |
| gampel1 | do we have a start up sequence problem or not ? | 12:04 |
| hshan | and reliability is based on that | 12:04 |
| gampel1 | i agree for only OVSDB updates | 12:04 |
| hshan | the startup sequence is solved in the patch | 12:04 |
| gsagie | hshan: are you also fixing L3 app ? | 12:05 |
| gsagie | as its using some part of the cookie | 12:05 |
| hshan | yes | 12:05 |
| gsagie | ok good :) | 12:05 |
| hshan | i add a the aging_cookie_bits to it's cookie | 12:05 |
| gsagie | ok so everything is good :) what is bad? | 12:06 |
| hshan | the bad is: | 12:06 |
| hshan | the fullstack test must run in serial | 12:06 |
| hshan | to make the restart of dragonflow don't affect other tests | 12:07 |
| hshan | can you accept it? | 12:07 |
| hshan | it will take long time than before | 12:07 |
| gsagie | hshan: maybe instead of actually restarting you can simulate a restart | 12:07 |
| gsagie | i think that one option we can do is add another job for tests that needs to be serialized | 12:08 |
| hshan | then it's not fullstack, maybe unit test, right? | 12:08 |
| gampel1 | I think that will be the best fullstack serial | 12:08 |
| hshan | I agree with your option | 12:08 |
| gsagie | hshan: ok, so convert to serial | 12:08 |
| gsagie | i will work on adding another job | 12:09 |
| gampel1 | please update in the patch that you are closing the startup seq | 12:09 |
| gsagie | for the fullstack tests | 12:09 |
| gampel1 | bug | 12:09 |
| hshan | okay | 12:09 |
| gampel1 | good then we can move only test that must run one by one there | 12:10 |
| gsagie | yes | 12:10 |
| hshan | cool! | 12:10 |
| gampel1 | hshan: any other issues | 12:10 |
| DuanKebo | cool, gal, i think about it but don't know how! | 12:10 |
| hshan | and the reliability feature rely on set-controller and check-controller-is-set | 12:10 |
| gsagie | DuanKebo: np i will do it, do you know how to temporary make the current tests run not in parallel? | 12:11 |
| gsagie | hshan: i will update the code to not delete the controller | 12:12 |
| gsagie | if its already set | 12:12 |
| oanson | Maybe move the test that rely on checking whether flows exist or not to the serial job as well. | 12:12 |
| hshan | current, it is controlled by testr's configuration | 12:12 |
| DuanKebo | yes i know | 12:12 |
| gsagie | ok cool | 12:12 |
| DuanKebo | Sg code is not very easy to review. do you need dingboot or yuanwei to introduce the code? | 12:12 |
| gampel1 | oanson: maybe after the stable tag is set | 12:12 |
| gsagie | DuanKebo: we need it to work :) i tested a simple case and it broke up with an exception | 12:13 |
| gsagie | lets make sure it works good and then we can optimize it together | 12:13 |
| DuanKebo | sorry to hear that. | 12:13 |
| oanson | gampel: It will make these tests more reliable | 12:13 |
| gsagie | its ok, np | 12:13 |
| gsagie | i added the exception i got and scenario on the review as a comment | 12:13 |
| oanson | Right now they fail for no reason other than other tests run at the same time. | 12:13 |
| hshan | Duankebo: we could provide another config for tesr in 'tools' | 12:13 |
| DuanKebo | I will ask yuanwei to help debug it | 12:13 |
| gampel1 | DuanKebo: for the selective proactive we need to fix the state that we get notify before the Port data is in teh DF DB | 12:14 |
| gsagie | if gampel wants we can also do a "code walk throught" on the code in another meeting maybe as you suggested | 12:14 |
| gampel1 | sound good | 12:15 |
| DuanKebo | Eran, have you encountered that case? | 12:15 |
| gsagie | hshan: just convert it to serial at this point and i will add the additional job that does it differently | 12:15 |
| gampel1 | yes when i ran fullstack locally | 12:15 |
| Fei | Anybody can verify the DNAT:) Hujie and me did the test, it's OK. Hope all of you can test it also. | 12:16 |
| hshan | ok, I've make it serial locally | 12:16 |
| DuanKebo | OK, I will work on it. | 12:16 |
| gampel1 | I will send you the log tomorrow | 12:16 |
| gampel1 | fei: i will test it tomorrow | 12:16 |
| Fei | @Eran, thank you | 12:17 |
| DuanKebo | I thought it won't occur at least currently. | 12:17 |
| gsagie | hshan: not locally.. i mean change testr configurations and send the patch :) | 12:17 |
| gampel1 | I run my devstack in a VM so maybe every thing is slower | 12:17 |
| hshan | gsagie: ok | 12:18 |
| gampel1 | DuanKebo: once the redis patch pass the CI I think we can merge | 12:18 |
| DuanKebo | OK, it's weird that the code can't pass ut. | 12:19 |
| nick-ma | seems a bug in unit test code. | 12:19 |
| gampel1 | yes | 12:19 |
| gampel1 | in the redis unit test | 12:19 |
| hshan | almost forgot, I delete a test from test_ryu_base_app.py | 12:20 |
| *** DuanKebo_ has joined #openstack-dragonflow | 12:20 | |
| gampel1 | why ? | 12:20 |
| hshan | we've decided to get all the update from OVSDB | 12:21 |
| gampel1 | hshan: Ok i see, you know you need to rebase your patch | 12:21 |
| hshan | and that test check features_reply_handler is called in all apps | 12:21 |
| hshan | ok, i will | 12:22 |
| gampel1 | hshan: where do you store the ID now in the Flow or file or DB | 12:23 |
| hshan | in the Flow | 12:23 |
| *** DuanKebo has quit IRC | 12:24 | |
| hshan | store in DB, if we lose connection with DB, the reliability cannot work | 12:24 |
| oanson | hshan: Same is true if we lose connection with vswitchd | 12:24 |
| gampel1 | Ok I meant OVSDB | 12:25 |
| oanson | When stored as a flow | 12:25 |
| *** DuanKebo has joined #openstack-dragonflow | 12:25 | |
| *** DuanKebo_ has quit IRC | 12:25 | |
| gampel1 | I think for this phase it is Ok we could consider changing that latter thoughts ? | 12:26 |
| gsagie | i agree | 12:26 |
| hshan | I agree | 12:26 |
| DuanKebo | sorry for offline. Agree what, heshan? | 12:27 |
| hshan | currently, choose to store cookie in flow | 12:27 |
| hshan | not in Central DB | 12:27 |
| gsagie | DuanKebo: talking about where to get the current cookie id for reliability (from flow or from DB either OVSDB or DF DB) | 12:27 |
| gsagie | hshan: you can also store it as gampel suggested in ovsdb | 12:28 |
| DuanKebo | ok, i see | 12:28 |
| gsagie | locally | 12:28 |
| gampel1 | Any other open issues ? | 12:28 |
| gsagie | not from me | 12:28 |
| *** yamamoto has joined #openstack-dragonflow | 12:28 | |
| DuanKebo | no | 12:28 |
| gampel1 | so we have feature freeze on the 7th | 12:28 |
| oanson | I have an issue in CI with the OVS service | 12:29 |
| gampel1 | and code freeze on the 18th | 12:29 |
| gampel1 | whats the issue ? | 12:29 |
| oanson | The tests fail to start. I don't know why. | 12:29 |
| hshan | gsagie: I'll think about it | 12:29 |
| oanson | I get "ValueError: '/usr/local/lib/python2.7/dist-packages' is not in list" | 12:29 |
| oanson | Anyone familiar with such an issue? | 12:30 |
| DuanKebo | Maybe you need to reinstall virtualenv? | 12:30 |
| Fei | which test case? | 12:31 |
| oanson | DuanKebo: Maybe, but I'm not sure how | 12:31 |
| DuanKebo | I have encountered similar error, found solution in "stackoverflow" | 12:31 |
| oanson | Fei: It doesn't start the tests. It fails on creating the virtualenv | 12:31 |
| Fei | ok | 12:31 |
| DuanKebo | I can send you the solution tomorrow. | 12:32 |
| DuanKebo | I'm not in office now. | 12:32 |
| oanson | DuanKebo: Thanks. I've been looking at this error for a few days now. | 12:32 |
| DuanKebo | Heshan, is yuanwei still in office? | 12:32 |
| oanson | I'd be happy to see what's the solution :) | 12:32 |
| gampel1 | Are there any other issues ? | 12:33 |
| Fei | maybe you can reinstall the pip | 12:33 |
| hshan | Yuanwei is not here | 12:34 |
| nick-ma | np. | 12:34 |
| gsagie | oanson try this: | 12:34 |
| gsagie | sudo pip uninstall virtualenv | 12:34 |
| gsagie | sudo pip install virtualenv==14.0.5 | 12:34 |
| gsagie | or upgrade virtualenv to 15.0.1 | 12:34 |
| oanson | gsagie, it's on the gate. On my environment it works | 12:34 |
| DuanKebo | yes, you can try gal's method | 12:34 |
| oanson | I'll try upgrading the gate to virtualenv 15.0.1 via the scripts. | 12:35 |
| hshan | Yuanwei is back | 12:35 |
| gsagie | oanson: link? | 12:35 |
| oanson | http://logs.openstack.org/97/296897/4/check/gate-dragonflow-dsvm-fullstack-nv/6f4999d/console.html | 12:35 |
| oanson | pip-freeze says the version is 14.0.6 | 12:35 |
| DuanKebo | devstack will install new viertualenv version 14.0.6 | 12:36 |
| oanson | Yes. | 12:37 |
| DuanKebo | If you run fullstack and devstack in the same machine, there will be a problem. | 12:37 |
| gampel1 | Ok thank very much everyone good evening | 12:39 |
| DuanKebo | Bye! | 12:40 |
| Fei | bye! | 12:40 |
| nick-ma | bye~ | 12:40 |
| gampel1 | bye | 12:40 |
| oanson | bye | 12:40 |
| hshan | bye | 12:40 |
| openstackgerrit | Li Ma proposed openstack/dragonflow: Delete stale df db objects when neutron db transaction failed https://review.openstack.org/297519 | 12:40 |
| gsagie | 晚安 | 12:41 |
| *** hujie11 has joined #openstack-dragonflow | 12:43 | |
| *** yamamoto has quit IRC | 12:44 | |
| *** yamamoto has joined #openstack-dragonflow | 12:47 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Install OVS as an OS service https://review.openstack.org/296897 | 12:48 |
| *** yamamoto has quit IRC | 12:52 | |
| *** yamamoto has joined #openstack-dragonflow | 12:55 | |
| openstackgerrit | hujie proposed openstack/dragonflow: Fix ovsdb monitor connect to ovsdb fail bug https://review.openstack.org/298167 | 12:56 |
| *** yamamoto has quit IRC | 13:12 | |
| *** DuanKebo has quit IRC | 13:15 | |
| *** yamamoto has joined #openstack-dragonflow | 13:17 | |
| *** yamamoto has quit IRC | 13:21 | |
| *** DuanKebo has joined #openstack-dragonflow | 13:23 | |
| *** yamamoto has joined #openstack-dragonflow | 13:47 | |
| *** openstackgerrit has quit IRC | 13:48 | |
| *** openstackgerrit has joined #openstack-dragonflow | 13:49 | |
| *** yamamoto has quit IRC | 13:52 | |
| *** DuanKebo has quit IRC | 13:52 | |
| *** yamamoto has joined #openstack-dragonflow | 14:12 | |
| *** yuli_s has quit IRC | 14:19 | |
| openstackgerrit | Omer Anson proposed openstack/dragonflow: Install OVS as an OS service https://review.openstack.org/296897 | 14:53 |
| *** oanson has quit IRC | 15:14 | |
| *** yamamoto has quit IRC | 15:16 | |
| *** yamamoto has joined #openstack-dragonflow | 15:17 | |
| *** yamamoto has quit IRC | 15:21 | |
| *** yamamoto has joined #openstack-dragonflow | 15:28 | |
| *** yamamoto has quit IRC | 15:28 | |
| *** yamamoto has joined #openstack-dragonflow | 16:28 | |
| *** yamamoto has quit IRC | 16:36 | |
| *** oanson has joined #openstack-dragonflow | 17:19 | |
| *** oanson has quit IRC | 18:08 | |
| *** gampel1 has quit IRC | 20:41 | |
| *** irenab_ has joined #openstack-dragonflow | 22:35 | |
| *** irenab has quit IRC | 22:38 | |
| *** irenab_ is now known as irenab | 22:38 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!